Reid Hoffman has expressed his opinion on the ‘tokenmaxing’ debate
Only days after Meta deactivated its internal “tokenmaxxing” dashboard—following the leak of its AI leaderboard to the media—LinkedIn co-founder and venture capitalist Reid Hoffman voiced his support for the concept. The “tokenmaxxing” trend has recently become a dominant focus across Silicon Valley.
An AI token is a basic unit of data that a model processes to interpret instructions and produce text. These pieces serve as the primary metric for measuring AI usage and calculating the cost of services.
Consequently, several organizations have started monitoring internal token consumption to identify which employees are most aggressively adopting AI tools. This trend has been dubbed “tokenmaxxing,” borrowing the Gen Z suffix “-maxxing”—a term for extreme optimization seen in other viral slang like “looksmaxxing” or “sleepmaxxing.”
Nevertheless, tech industry engineers are debating whether this metric serves as a valid proxy for productivity, with many arguing that it is effectively equivalent to ranking employees based on who spends the most money.
Speaking at Semafor’s World Economy Summit, Reid Hoffman shared a favorable view of companies monitoring AI usage. While he didn’t use the slang term “tokenmaxxing,” he explicitly recommended that businesses track how many tokens their employees are spending as they integrate AI.
“You should be encouraging people across every function to actually engage and experiment [with AI],” Hoffman remarked at the event. He suggested that while not a perfect indicator of productivity, monitoring token usage serves as a valuable dashboard for leaders to gauge real-world participation.
He clarified that high token consumption can sometimes stem from random or purely exploratory use. Consequently, he advised that “tokenmaxxing” data should be contextualized by analyzing the specific tasks and objectives employees are pursuing with those tokens.
Hoffman noted that while many experiments will inevitably fail, that process is essential. He emphasized the importance of maintaining that feedback loop, advocating for a broad, diverse group of employees to engage with the technology “collectively and simultaneously.”
Beyond tracking usage, Hoffman advised companies to embed AI across their entire organization. He also stressed the importance of holding regular meetings where employees can exchange insights and discuss which AI workflows are delivering the best results.
Hoffman recommended instituting a weekly rhythm for reflection, suggesting that teams gather to discuss how they’ve applied AI toward personal and organizational productivity. “What you’ll find is that some of the discoveries are truly remarkable,” he noted, emphasizing that the goal is to share specific lessons learned from new experiments.
